O'Reilly logo
  • Chuck Deatherage thinks this is interesting:

Automating DQAF measurement requires building functions to collect raw measurement data, execute appropriate calculations against that data, and compare against thresholds or past history to make the data meaningful so that data’s quality can be assessed. The ways of collecting raw measurement data can be understood largely in terms of what features of the data need to be counted (file sizes, process duration, record counts, group-bys, etc.). Making measurements meaningful can be understood through the kinds of calculations (percentages, averages, etc.) that put results from large datasets into comprehensible terms and set up a means of making useful comparisons. The comparisons utilize statistical method...